Complex media such as biological tissue have complicated anisotropic physical properties (e.g. stiffness tensor, Willis coupling tensors) that depend on direction and are important to measure. For example, they may inform on the tissue health. We probe complex media with mechanical waves and process the scattered fields to learn the media dynamics using an array of methods including physics-based machine learning algorithms. This research has applications in the non-destructive evaluation of media, the design and characterization of metamaterials, and healthcare.
Marine mammals such as dolphins are excellent at using ultrasound to discover and navigate their underwater environment. Their biosonar is significantly more advanced than any human-made sonar. We look at dolphins for inspiration to understand how to efficiently use ultrasound to map the environment and classify the objects within. We use this knowledge to create new physics- and machine learning-based sensing solutions for autonomous vehicles and acoustic imaging devices such as sonar and medical ultrasound scanners.
Metamaterials – materials with carefully controlled microstructure – enable the manipulation of physical waves in new and unconventional ways. We use machine learning methods to design metamaterials with embed active elements (e.g., analog and digital electronic circuits, microcontrollers, etc.) for new generations of programmable, smart materials with functionalities and physical properties going beyond traditional media. These active media find impactful applications in many areas including sensing, imaging, communications, and healthcare.
Dr Popa obtained his PhD in Electrical Engineering from Duke University. After graduation he continued working as a Postdoctoral and Research Scientist on electromagnetic and acoustic metamaterials. He was a visiting professor at University of Bordeaux (2016) and joined the Mechanical Engineering department of the University of Michigan the following year.